This paper focuses on improvements of logistics via removal of weakness points inside the district Agro Nocerino Sarnese (Italy, region Campania), famous for the export of products of high quality to other Italian regions and/or foreign countries. The logistic chain is guaranteed by fleets of trucks, whose aim is to transport goods from a point to another inside the region. In order to repair eventual faults of trucks, the logistic network has particular nodes that are small family business, which have decision processes of naïf type dealing with tradition and related to a leadership. In absence of such leadership, a management crisis might occur, with consequent delays on repair processes and on the delivery of products to the final customers. Hence, possible improvements on performances of the logistic network are achieved trying to reproduce the leadership's decisions (making them stable) via a profiling/competencebased approach, enriched by experience. Precisely, models, dealing with profiles of clients, competences of workers and experience of the leadership, are mixed up and used to reconstruct the leadership's decisions. A real case of small business is useful to test the proposed approach.
Phenomena connected to logistics inside Italian regions are nowadays object of particular attention, due to the survival of regional economy as well as the wealth of the Italy itself. Indeed, region Campania represents a geographical portion in which some aspects connected to transports and distribution of products are fundamental in order to guarantee profits. This is evident thinking of various goods that region Campania produces. For instance, the famous pizza “di Napoli” is obtained via some operations that derive from high quality products, such as the tomatoes of San Marzano and/or the water of sources near the Vesuvius. It is obvious that other parts of Italy, as well as other foreign countries, require products of region Campania, with the aim of reproducing foods and meals of acceptable standards.
On the other hand, region Campania has the necessity of exporting its own products, as the primary intention is an increment of prestige at national and international level. For this reason, various efforts are made in order to guarantee efficient supply chains, where the satisfaction of producers meets the one of consumers, also thanks to an efficient logistic network that involves the whole region Campania. In particular, the so-called district “Agro Nocerino Sarnese” (ANS), namely an area in Campania, is useful to understand the complex dynamics of region’s logistics. In fact, the district ANS is characterized by the production of many consumer goods, that are then distributed throughout the region and exported to other parts of Italy and/or to foreign countries. One factor that implies wealth in region Campania is the quality of goods that are produced by the district. In particular, the production and the consequent export are the key factor for the success of all operations, which occur inside the region.
In particular, the transport of products from a point to another is often guaranteed by “padroncini”, i.e. owners of trucks. Padroncini aim at managing their fleets of vehicles in order to achieve high quality standards in terms of relationships of type producer-consumer. Indeed, extraordinary events, such as malfunctions and technical problems of trucks, might cause failures for the products’ transports, with consequent disadvantages for the final customer and the performances of the overall logistic network of region Campania.
Indeed, the logistic network has some particular nodes, consisting of small family
business, whose job is to repair possible faults of trucks. It follows that a possible
high critical situation for logistics inside region Campania depends on how activities
of small family business are managed, see [9]. In fact, delays in repairing faults are
easily translated into delays in distribution and consequent deterioration of products.
As a consequence, nodes, which represent small family businesses, influence the
logistic chain. Hence, a possible policy for improvements foresees the reproduction of
leadership’s decisions for the work teams assigned at each work order. In fact, the
leadership, thanks to its experience, knows the client that has to be served first and
how to serve him, see [
In this paper, with the aim of improving the logistics inside region Campania, we study a possible methodology for capturing the leadership’s decisions of family businesses. Such approach foresees the fusion of models that are often used individually or in pipe and deal with profiles, competences and experience, i.e. it is a profiling/competence-based approach, shaded with experience.
In particular, as for profiles, the clients’ characteristics are studied in order to
establish some priorities among the work orders that arrive at a generic small
business, see [
The approach is tested on a real small business inside region Campania. The results for work teams of first choice (the teams chosen by the leadership) and second choice (the alternative teams if first choices are not possible) are very positive. In particular, for the work teams of second choice, the proposed approach indicates that 90% of leadership’s decision is reproduced (also if the leadership is absent, and this makes the process stable).
The paper is organized as follows. In Section 2 the real conditions of logistics inside region Campania are described, with emphasis on the transport of goods and problems due to unexpected events. Section 3 considers the case study of a real critical node in region Campania. Section 4 reports the final research results. The paper ends with Conclusions in Section 5. 2
Inside region Campania, Italy, there is the area of Agro Nocerino – Sarnese (usually called district ANS), which consists of two different geographical parts, the Agro Sarnese and the Agro Nocerino, both located in the valley of the river Sarno, precisely between Naples and Salerno. The district ANS produces many consumer goods, distributed throughout the region and exported to other parts of Italy. The main groups of these products are the following: meat, fish, fruit, milk and dairy products, vegetables (in particular, tomatoes and onions).
Obviously, depending on the type, the previous goods undergo several steps in the producer-consumer chain. Their transit to the consumer is due to a dense logistic network, which involves the whole region Campania. In this context, many factors allow the success of the operations. In particular, we consider the speed of transportation from one point to another in the region, the need to maintain appropriate conditions on the goods during transports (for example, the existence of the cold chain for fish), the partial reduction of unexpected events, such as faults to the means that allow transits in the various points of the network. Notice that, as for big distances and quantities, large transport companies with internal workshops manage trucks that provide the location of goods to another node of the logistics network, thus ensuring an effective, efficient and redundant service. On the contrary, the whole universe of small and medium retailers, served primarily by “padroncini”, deals with the regional network.
These small transportation companies, often consisting of just one person, are fundamental to ensure the supply of products to small and medium-sized municipalities as they represent the backbone of the region. Padroncini use a number of workshops/establishments, which are inside the district ANS and are specialized in repair, customization and maintenance of trucks. An example of workshop is the Santonicola F.lli SNC, located in Siano, province of Salerno.
The establishments manage the padroncini’s demands because means useful for the transports of goods are often subject to malfunctions of different types, such as mechanical failures for the transport itself and technical problems relating to the storage of the product. The latter is much more critical because the necessary condition that guarantees intact products inside distribution centres is that conservation devices of trucks could work in optimal conditions. Considering such phenomena, as for the transits of trucks in the logistic network, there are further nodes, consisting of sets of small family businesses, whose task is to repair eventual faults of trucks. This suggests that a possible optimization or, in better terms, a possible improvement of the internal dynamics of the district ANS may depend on the management of activities inside small family business. Such situation is easy to understand if you think that delays in repairs of faults imply cascade delays for the distribution of products inside distribution centres, with consequent deterioration of products before they arrive at the final consumer. Hence, nodes representing small business are critical, as they influence the logistic distribution inside region Campania.
In order to overcome such problems, it is necessary to focus on these nodes and their activities. This subject is complex because the decision criteria within small businesses are very empirical, based on experience and tradition and taken by a leadership that, over the years, may be subject to changes. The transition from a leadership to the next one may cause negative trends, as possible decisions are not structured and the whole knowledge of the previous leadership is not transferred to the new generation. This has meaningful impacts on either served customers (in this case, padroncini) or the company itself. Such situations imply, undoubtedly, negative performances inside the logistic chain.
Considering the scenario of the district ANS, the aim of this analysis is to study a possible critical node, providing for it an appropriate solution to address the needs of padroncini and ensuring, therefore, a possible improvement of performances inside the logistic network. 3
In what follows, we consider the analysis of a real critical node of the logistic network inside region Campania. The activities of such enterprise are managed by a family and deal with single work orders that are satisfied according to clients’ exigencies.
In particular, the leadership considers each work order and, according to the clients’ types and priority reasons, assignes a work team, which consists of a subset of workers, the most suitable either for their skills/abilities or for the characteristics of work orders and priorities of padroncini. The enterprise has different workers, who are listed as follows and whose names are not reported for privacy. There are four coach builders (C1, C2, C3 and C4), four varnishers (V1, V2, V3 and V4), four welders (W1, W2, W3 and W4), two electricians (E1 and E2), three mechanics (M1, M2 and M3) and two upholsterers (U1 and U2).
We consider sixteen work orders to which the leadership assigned a work team. Unlike the case reported in [9], a deeper analysis allowed establishing the so-called teams of “second choice”, i. e. the teams chosen by the leadership when some of the most suitable workers for a given work order are already busy or not available. The analysed work orders and the work teams of second choice are in Table 1.
Notice that the work teams of second choice are not always used by the leadership. Indeed, for some particular work orders (1, 9, 12, 14, 15 and 16, indicated in gray in Table 1), the enterprise does not prefer to assign alternative teams because of the delicacy of the required actions. In these cases, the work teams are the first choices of the leadership. Hence, if it is not possible to assign the team of “first choice”, the work order is queued and the corresponding operations are made in a second moment. Moreover, queues of work orders might also occur if it is not possible to create work teams of second choice. In this last case, the work orders are completed with a priority, which is strictly dependent on the clients’ characteristics.
In order to improve all operations inside the logistic network, an obvious analysis of work orders and teams is necessary, with the aim of reproducing the possible choices of the leadership. This problem is very difficult, as it deals with the tradition and history of the enterprise. For this reason, various approaches, whose combination might imply a partial reproduction of the leadership’s choices, should be considered. Decomposing the original problem into possible factors that are suitable to find a correct solution, it is possible to recognize that the choices’ characteristics obey the following criteria: client’s features; competences of workers involved in work teams; leadership’s experience. Hence, the approach that we are going to follow is profiling/competence – based and it is enriched with experience. Indeed, there are many models dealing with such problems, but are often used separately or in pipe. The real efforts in this work foresee the fusion of existing approaches, with the aim of obtaining an acceptable reconstruction of leadership’s decisions.
In detail, as for the profiling, we follow a model based on convex combinations of factors (see [10] for an example) that are useful to represent the enterprise’s clients.
As for the competence representation, we consider Knowledge, Skills and
Attitudes, see [18] for details. In particular, Knowledge is the set of support information
for a given task; Skill is the capacity of developing the task; Attitude indicates a
particular behavior in facing some situations. Indeed, we indicate as KSA Model (see [
Finally, as for the leadership’s experience, Pattern Mining techniques are useful to
find relevant patterns in data sequences. In this paper, we use a Pattern Mining
algorithm called “Apriori” (see [
In what follows, first we describe some features of the chosen methods for profiling, competences and experience; then, we briefly analyse the approach for the reconstruction of leadership’s decisions.
Suppose that each client Ci , i = 1,..., KC , of the enterprise is indicated by a vector, 3.1
Profile modelling whose components are the following: • Sales figures, Ci1 . • • • •
For the client Ci , i = 1,..., KC , the component Ci1 is given by the average sale figures in last three years, while the component Ci3 indicates the average number of work orders in last three years. The remaining parameters Ci2 , Ci4 and Ci5 are variable and generally estimated by the enterprise.
The priority of each client Ci , i = 1,..., KC , is computed as:
5 pCi = ∑ω jCij ,
j=1 5 where 0 <ω j < 1, j = 1,..., 5, ∑ω j = 1.
j=1 3.2
For the analysed enterprise, two different types of KSA models are considered, precisely for workers and work orders. Notice that a different KSA model is defined for each type of worker (coach builder, electrician, mechanic, upholsterer, varnisher and welder). Then, following the characteristics of the various work orders, KSA models for work orders are constructed. Indeed, KSA models for workers and work orders have a precise difference: the former focus on all characteristics of workers while the latter consider what is useful for an assigned work order. For a better comprehension, consider Table 2, that reports the thirtytwo concepts of the KSA model for the worker “varnisher”, divided into Knowledge, Skills and Attitudes. In particular, the eight bold concepts allow constructing the KSA model for the work order 3.
Notice that the concepts for each KSA model (either for workers or for work orders) correspond to different skos:ConceptScheme and skos:Concept, as shown in Fig. 1, that represents an extract of the KSA Model for Welders (shortly indicated by KMW).
From Fig. 1, we get that the category Knowledge has the subcategories Mechanics and Electronics; Skills has treatment of materials and use of devices, and so on. The ellipses indicate the subsets of welder qualities useful to construct a Worker Order KSA model (shortly indicated by WOKMW), and precisely the work order 9. Notice that, using the only KMW, the welder to choose is W2 (the best worker for the work order), while WOKMW proposes W4. This last choice is coherent with the leadership one (see Table 1), as it indicates the best worker for the assigned work order.
The interaction among workers is an important parameter for the success of an enterprise. Indeed, the leadership often foresees that some workers have a high degree of cooperation. Such phenomenon is difficult to model, as it deals with the behavioural attitudes of single workers and the empathy with other colleagues. This situation is captured via the “Apriori” algorithm. It considers that, if a given item set is frequent, its subsets are frequent too. Precisely, consider a transaction database T and a support threshold !. Let !! be the candidate item set of length k and !! be the frequent item set of length k. A possible pseudo code for the algorithm is the following: Apriori (T,ε ) L1 ← {itemset of length 1} k ← 2 while Lk−1 ≠ ∅ Ck ← generate Lk−1 for transactions t ∈T , Ct ← subset (Ck ,t ) ; for candidates c ∈Ct , count [c] ← count [c] +1; Lk ← {c ∈ Ck : count [c] ≥ ε } ; k ← k +1 return U Lk
k
The Apriori algorithm was able to identify the following couples of workers that often occur in work teams of second choices: (C2, V1), (V1, W2), (V3, W3), (E1, M1) and (V1, W4). Such couples represent persons who often work together, and depend only on the leadership itself, i.e. on its experience. The found couples are a quite meaningful example to capture the leadership’s experience, a heritage that is difficult to transmit to future generations. 3.4
The whole approach foresees the fusion of the three chosen models (often used individually, see subsections 3.1, 3.2 and 3.3) in order to capture suitable work teams of first and second choice. The steps of the approach are the following (consider Fig. 2 for an abstract vision): • Reorder the work orders according to the clients’ importance and priorities. • For each work order, choose the team obtained via the KSA models. If such teams coincides with some choices of the leadership, stop; otherwise, try to change workers of work teams using the association rules obtained by the Apriori algorithm. • If, for the work order in consideration, first and second choice’s teams have already been assigned, put the client in queue.
In this section, we present the results obtained for the enterprise described in Section 3. In particular, Table 3 reports the comparison between the leadership’s work teams (LWT) of first and second choice for the various work orders (WOs) and the ones obtained via the fusion among profiling modelling (clients’ characteristics), competence modelling (KSA models), and experience (APriori algorithm). Gray lines indicate cases in which the leadership does not foresee second choices, and the reconstruction’s percentage is also indicated for each reconstructed team.
WOs As for the first choices, the approach is able to capture a total correspondence for the work orders 1, 2, 3, 4, 6, 7, 8, 9 and 16, i.e. a 100% reconstruction occur for 9/16 situations. As for second choices, considering that work orders 1, 9, 12, 14, 15 and 16 have not to be analyzed, we have a 100% correspondence in cases 2, 3, 4, 5, 6, 7, 8, 11 and 13.
Such results are in Table 4 in terms of correspondence percentages.
For the work teams of first choice, 13/16 work orders (about 80%) present a more than 50% correspondence with the leadership’s decisions. • For the work teams of second choice, 9/10 work orders (90%) have a more than 50% correspondence with the leadership’s decisions.
This indicates that the proposed approach is quite useful for the reconstruction of the enterprise’s activity with a high degree of accuracy.
10 8 6 4 2 0 100%
More than 50 %
Less or equal than 50%
In this paper, an analysis of logistics inside the district Agro Nocerino Sarnese (Italy, region Campania) is studied. Possible improvements are obtained via the reproduction of decisions inside small family businesses.
Considering a real case of enterprise and using a profiling/competence-based approach, enriched with experience, it was possible to reproduce the dynamics of leadership’s choices with a 80% and 90% degree accuracy, respectively, for work teams of first and second choice. 6 Platform, In: Web Information Systems Engineering – WISE 2011 and 2012 Workshops, pp. 133 – 146 (2013), Springer 8. Gaeta, A., Gaeta, M., Piciocchi, P., Ritrovato, P., Vollero, A.: Evaluation of the human resources relevance in organisations via knowledge technologies and semantic social network analysis. Internation Journal of Knowledge and Learning. In press (2014) 9. Gaeta, M., Perano, M., Piciocchi, P., Rarità, L.: Construction of decision criteria for family – run business, In: Proceedings itAIS 2014, pp. 1 – 12 (2014) 10. Gaeta M., Piscopo R., Rarità L., Trevisant L., Novi D.: A Knowledge Management Strategy to Identify an Expert in Enterprise. Smart Organizations and Smart Artifacts, Lecture Notes in Information Systems and Organisation, Springer International Publishing Switzerland 2014. 7, pp. 173 – 182 (2014) 11. Gu, J., Wang, B., Zhang, F., Wang, W., Gao, M.: An Improved Apriori Algorithm. Communications in Computer and Information Science. 224, pp. 127 – 133 (2011) 12. Hong, T., Kuo, C., Wang, S.: A fuzzy AprioriTid mining algorithm with reduced computational time. Applied Soft Computing. 5(1), pp. 1 – 10 (2004) 13. Hong, T. P., Wang C. Y., Lin C. W.: Providing timely updated sequential patterns in decision making. International Journal of Information Technology & Decision Making. 9(6), pp. 873 – 888 (2010) 14. Jia, C. Y., Ni, X. J.: Association rule mining: A Survey. 30, pp. 145 – 149 (2011) 15. Kaur, C.: Association Rule Mining using Apriori Algorithm: A Survey. International Journal of Advanced Research in Computer Engineering & Technology. 2(6), pp. 2081 – 2084 (2013) 16. Lin, C., Hong, T., Lan, G., Wong, J., Lin, W.: Incrementally mining high utility patterns based on pre – large concept. Applied Intelligence. 40(2), pp. 343 – 357 (2014) 17. Loia, V.: Special issue on new trends for ontology – based knowledge discovery.
International Journal of Intelligent Systems. 25(12), pp. 1141 – 1142 (2010) 18. Loia, V., De Maio, C., Fenza, G., Orciuoli, F., Senatore, S.: An enhanced approach to improve enterprise competency management, In: 2010 IEEE World Congress on Computational Intelligence, WCCI (2010) 19. Patel, B., Chaudhari, V. K., Karan R., K., Rana YK: Optimization of Association Rule Mining Apriori Algorithm Using ACO. International Journal of Soft Computing and Engineering. 1(1), pp. 24 – 26 (2011) 20. Sampson, D., Fytros, D.: Competence models in technology – enhanced competence based learning, In: H. H. Adelsberger, Kinshuk, J. M. Pawlowski, D. G. Sampson (Eds.), Handbook on Information Technologies for Education and Training, International Handbooks on Information Systems, Springer Berlin Heidelberg, pp. 155 – 167 (2008), doi:10.1007/978-3-540-74155-8-9 21. Sumithra, R., Paul, S.: Using distributed apriori association rule and classical apriori mining algorithms for grid based knowledge discovery, In: 2010 International Conference on Computing Communication and Networking Technologies, pp. 1 – 5 (2010)